Professional Documents
Culture Documents
Structure
9.0 Introduction
9.1 Learning Outcomes
9.2 Nature of Experimental Method
9.3 Classic Experimental Research Design
9.4 Process of Experimental Research
9.4.1 Control of Extraneous Variables
9.4.2 Randomisation
9.5 Experimental Design
9.5.1 Factorial studies
9.5.2 Quasi experimental design
9.6 Field Experiments
9.7 Merits and Demerits of Experimental Method
9.8 Let Us Sum Up
9.9 Keywords
9.10 References and Further Readings
9.11 Check Your Progress: Possible Answers
9.0 INTRODUCTION
Communication research has established itself on a firm footing and is a fertile
field for young researchers. The area is rich, vast and varied. Experimental method
is a well-established method of media research but its use is restricted. There is
a tendency among researchers to opt for the survey method. Experimental research
is highly popular in psychology and social psychology. In the case of media
research, experimental method can produce an enormous amount of data.
However, media researchers find several issues for conducting experiments. For
instance, the method is largely quantitative and uses extensive statistical
procedures for analysis. Issues relating to infrastructure, i.e. getting the appropriate
testing conditions, and/or getting matching subjects for experiments pose
problems for researchers. There can be laboratory experiments or field
experiments. The purpose is to find out and establish a causal relationship between
an independent variable and a dependent variable. This method is also amenable
to statistical analysis depending on the data generated.
We have discussed earlier that the selection of a research method depends upon
the topic selected and the researcher’s methodological approach towards the
selected topic. The experimental technique supports empirical approach to
research.
122
Experimental Methods
9.3 CLASSIC EXPERIMENTAL RESEARCH
DESIGN
A classic experimental research design contains two groups for comparison. One
is an experimental group and the other is a control group. These two groups are
equal in every respect. However, there is one difference. The experimental group
is administered or exposed to an independent variable, which is called
experimental stimulus, and the control group is not exposed to the experimental
stimulus. Measurements on the dependent variable are called scores and are taken
twice from each group. The first measurement is known as the pre-test. It is
taken prior to the administration of experimental stimulus to the experimental
group. The control group also undergoes measurement. The second set of
measurement takes place after the experimental group is exposed to the
experimental stimulus. This is known as the post-test. The control group, of
course, is not exposed to the experimental stimulus. But both the groups are
measured. The difference in measurement between the post-test and pre-test will
be compared in each of the two groups, experimental and control. In case of the
difference in the experimental group, we can safely assume that the independent
variable causes changes in the dependent variable, or the independent variable
has an effect on the dependent variable and is causally related dependent variable.
Group Pre-test Post-test Difference
Experimental Ml M2 M2-Ml =Ra
Control M3 M4 M4- M3 =Rb
Selden Menefee and Audrey Menefee who visited Tiptur area of Karnataka
conducted an experimental study using a mimeographed newspaper to measure
the information level of sample subjects chosen. When the measurements were
compared between the experimental group and the control group, it was found
that the information level of the experimental group had gone up due to the
exposure to the mimeographed newspaper more than the control group that was
not exposed to the newspaper, the experimental stimulus. This meant that exposure
to the newspaper caused the increase in the information level of those exposed to
it.
Below is a direct quotation from Wimmer and Dominick (2003) who have
identified the steps as:
1) “Every researcher has to select the setting. Most experiments are conducted
either in a laboratory or any other environment which will be under the
control of the investigator. Experiments are also conducted under natural
settings, but the researcher will have limited control over his experiment.
2) Any research will have to select the experimental design. The hypotheses
or research question will determine the appropriateness of the design. The
types of variable available for manipulation or evaluation, availability of
respondents, and resources will impact the design selection.
4) Decide how to manipulate the independent variable. A set of clear and specific
instructions, events or stimuli is developed to manipulate the independent
variables and present them to the subjects of the experiment. The two types
of manipulations are straight forward and staged. “In straight forward
manipulation, written materials, verbal instructions, or other stimuli are
presented to the subjects”. If it is staged manipulation, the investigators
create events and circumstances that help them to manipulate the independent
variable. It can be simple or elaborate. The use of a confederate is also in
vogue. A confederate is a person who acts like a subject, but in reality, he is
also part of the manipulation.
The random assignment of subjects to the experimental and control groups will
ensure equality between the groups and also internal validity of the investigation.
However, it may not be representative of the larger population, but represent
only the sample selected. High rate of refusal and enormous cost may affect
opting for a highly representative sample. Next will be the reactive arrangements.
The results of a study should not only be generalised to a large population, but
also to a real-life situation. This is difficult to accomplish. It may result in reactive
effect of testing on the subjects. For instance, a pre-test may affect the response
of subjects to the experimental stimulus. Such an effect can be avoided by opting
for post-test only.
129
Research Methods – There are a few practical difficulties with the field experiments. For instance,
Quantitative
certain research studies cannot be done in the field due to ethical reasons. The
impact of television violence on children cannot be measured through field
experiments. Field experiments now and then face external obstacles that may
be unexpected. It is time-consuming when the researcher wants to establish
contacts, seek co-operation and obtain necessary permission. Often, this procedure
takes weeks and months. The most difficult part of field experiment is the control
of intervening variables. The extraneous variables will affect the accuracy of
experiments, and in turn the entire research work.
130
Disadvantages: Experimental Methods
A classic research design contains two equal groups: experimental and control.
The experimental group is exposed to experimental stimulus and the control
group is not. The measurements taken during pre-test and post-test are compared
to know whether the experiment involves manipulation and control, which are
difficult to achieve in social sciences. The experimental technique will normally
have a process of eight steps. Control of confounding or extraneous variables
can be done by matching, randomisation and generalisation. Internal and external
validity will have to be achieved in experimental research. There are two designs
when we deal with experimental method: Experimental design and quasi
experimental designs pretest-posttest control group, post-test only group and
Solomon four-group designs belong to experimental design.
Experimental designs are not suitable for application in certain research situations,
especially when manipulation and random assignment of respondents are not
possible. Experimental designs are best suited to stimulus-response relationships
but not for property-disposition relationships. The property-disposition
relationships can be tackled by quasi-experimental designs that include one-shot
case study, correlational designs, panel and trend studies. Field experiments are
conducted in natural settings. They are closer to reality and often non-reactive.
They do not need any special equipment. The costs can be less compared to
other methods. The advantages of experimental methods are evidence of causality,
control, cost benefit and replication. The disadvantages are artificiality,
researcher’s bias and limited scope.
9.9 KEYWORDS
Confounding variables: Variables which are extraneous to the experiment. They
contaminate the findings.
Deductive explanation: An explanation that has the element of generalisation
regarding past, present and future.
Empiricism: Verification with strong evidence.
Explicandum: An event or phenomenon to be explained.
132
Manipulation: Deliberately produce a change in the phenomenon selected. Experimental Methods
Pilot study: Involves a small number of subjects on whom the study is conducted.
Subjects: People who are selected to participated in the experiment.
Verstehen: A tradition of understanding human behaviour.
134